Case Study: Sentiment Text Analysis

The client “TasNetworks is a Tasmanian state-owned corporation that supplies power from the generation source to homes and businesses through a network of transmission towers, substations and powerlines.” The objective To uncover the sentiment and topic clustering hidden within the verbatim responses collected by the monthly ‘Satisfaction’ and ‘Ease Of Doing Business’ survey over a twelve month period. The customer verbatim sentiment analysis The textual data for each response was analysed for sentiment using a sentiment model which identifies the positive/negative/neutral polarity in the textual communication. The local polarity of the different sentences in the text is identified, and the relationship between them evaluated, resulting in a global polarity value for the whole text for a participant’s comment. This was done on a comment by comment level and scores were aggregated to provide a monthly or total score. The customer verbatim comment topic clustering We identified the text clusters by grouping a set of texts in such a way that comments in the same group (called a cluster) are more similar to each other than to those in other clusters. The clustering text algorithm receives a set of comments and returns the list of detected clusters. Each cluster is assigned a descriptive (topic name), a relevance value (indicating the relative importance of the cluster when compared to all other clusters), how often this descriptive topic is found, and the list of text elements that are included in the cluster.  Each comment may be assigned to one or several clusters. We then identified the scale and size of the sentiment and clusters based on the frequency and score. This provided...